Correlation and Path Coefficient Analysis for Seed Yield and Its Component Traits in Indian Mustard (Brassica juncea L.) under Western Uttar Pradesh Conditions

Introduction

Among the major oilseed crops grown across India, Indian mustard (Brassica juncea L.) holds a prominent place, especially in the northern plains, where it is vital for ensuring the country's edible oil supply. This crop accounts for a large share of total oilseed output and provides a primary livelihood source for farmers, particularly in areas like Western Uttar Pradesh. Recent figures show that rapeseed-mustard covers about 8.6 to 9.3 million hectares in India, with total production estimated between 12.6 and 13.2 million metric tons during the 2023-24 and 2024-25 growing seasons (USDA-FAS, 2024). Despite some gains in recent years, India's average crop productivity, roughly 1440-1460 kg per hectare, still falls short of the worldwide average (FAOSTAT, 2023). On a global scale, rapeseed-mustard is planted on around 44 million hectares, yielding close to 92 million tonnes, with an average productivity of about 1980-2070 kg per hectare (FAOSTAT, 2023). This indicates that India achieves only about 60-70% of the world average productivity, highlighting a substantial yield gap and emphasizing the need for genetic improvement and efficient breeding strategies (Kumar et al., 2025) [7] . Even though Indian mustard holds significant economic value, its yield continues to stay below achievable levels. This shortfall is largely attributed to the complicated manner in which seed yield is inherited and its reliance on many mutually influencing characteristics (Gupta et al., 2025) [6] . Seed yield in this crop is a quantitative attribute shaped by various morphological and physiological factors, such as plant stature, primary and secondary branch numbers, siliqua count per plant, seeds inside each siliqua, and seed weight. These components often show varying levels of correlation both among themselves and with final seed yield, making direct selection less effective (Kaur et al., 2025) [7] . Similar observations have emerged from broader Brassica improvement studies, where yield is determined by intricate interactions between component traits and environmental influences, thereby requiring the application of advanced biometrical tools for efficient selection. Correlation analysis helps estimate how strongly and in what direction yield is associated with its contributing traits. However, this method does not separate direct effects from indirect ones. Path coefficient analysis fills this gap by partitioning correlation coefficients into direct and indirect components, offering a clearer view of cause-and-effect connections among traits (Devi et al., 2025) [2] . Research conducted across different environments has shown that traits like biological yield, harvest index, and number of siliquae per plant have strong direct impacts on seed yield, confirming their value as dependable selection criteria (Kumar et al., 2025) [7] . Moreover, the expression of both yield and its components is heavily affected by environmental conditions. Consequently, evaluating genotypes in a region-specific manner is crucial to identify stable, high-yielding traits suited to particular agro-climatic conditions. In Western Uttar Pradesh, where environmental fluctuations strongly influence crop performance, there is limited knowledge about how yield components interrelate in Indian mustard (Gadi et al., 2025) [5] . Therefore, the current study was carried out to examine the correlation and path coefficient relationships between seed yield and its component traits, aiming to determine effective selection criteria for yield improvement under the conditions prevailing in Western Uttar Pradesh.

Materials and Methods

Site Description and Soil Properties

The field experiment was carried out at the research farm attached to C.C.R.D. College in Muzaffarnagar during the rabi (winter) season. The farm lies at 29°45' N latitude, 77°70' E longitude, at an elevation of 272 meters above sea level. This area falls within a subtropical climate zone, marked by a distinct winter season (October to March) with cool conditions (average minimum 6-7°C, average maximum 21-23°C) and occasional rain brought by western disturbances. The soil at the experimental site was sandy loam in texture, alkaline in reaction, low in organic carbon, and medium in available nitrogen, phosphorus, and potassium. It was well-drained and had moderate waterholding capacity.

Genetic Material and Experimental Layout

Ten Indian mustard accessions, used as parents, were obtained from the germplasm collection maintained at NBPGR, New Delhi. These were: Swarna, TERIGZ05, IC-491262, IC-491403, IC-491520, IC-491560, IC-520764, IC-521378, IC-546946, and Divya. These ten diverse genotypes were crossed in a diallel mating design (excluding reciprocals) during the rabi season of 2022-2023. The final experimental set included 55 treatment combinations (10 parents + 45 F₁ hybrids). The trial was arranged in a randomized block design (RBD) with three replications during the rabi season of 2023-2024. Each parent and F₁ hybrid was planted in a single row of 5 meters length, with rows spaced 45 cm apart. Within each row, plants were thinned to maintain a 15 cm spacing between plants. Standard agronomic practices were followed to ensure healthy crop growth. A uniform fertilizer dose of 60 kg nitrogen, 40 kg phosphorus pentoxide (P₂O₅), and 20 kg potassium oxide (K₂O) per hectare was applied at sowing. Irrigation and plant protection measures were carried out as needed.

Traits Measured

From each genotype within every replication, five competitive plants were chosen randomly to record data on the following quantitative characteristics: days to first flower (DFF), days to 50% flowering (DF50), days to maturity (DM), primary branches per plant (PB), secondary branches per plant (SB), plant height (PH), main raceme length (MRL), leaf area index (LAI), siliquae per plant (SP), seeds per siliqua (SS), thousand-seed weight (TSW), biological yield per plant (BY), seed yield per plant (SY), harvest index (HI), and oil content (OC).

Statistical Methods

Genotypic and phenotypic correlation coefficients among all pairs of traits were computed following standard methods. Path coefficient analysis was carried out using seed yield per plant as dependent (effect) variable, with other traits as independent (cause) variables, to partition correlations into direct and indirect effects. All the statistical computation were performed using standard biometrical procedures suggested by Searle (1961) [20] and Dewey and Lu (1959) [3] respectively.

Correlation Analysis

To understand relationships among traits, both phenotypic and genotypic correlation coefficients were calculated. Phenotypic correlation shows the observable association between characters, while genotypic correlation reflects the underlying genetic relationship after removing environmental effects. Comparing both levels helps clarify how traits depend on each other and guides selection decisions. For any two traits, the correlation coefficient was derived by dividing their covariance by the square root of the product of their individual variances. Variance and covariance components were obtained from the analysis of variance and covariance tables. A t-test was then used to check whether the resulting correlations were statistically meaningful or simply due to chance.

Path Coefficient Analysis

Because correlation alone cannot tell whether a trait directly influences yield or acts through other traits, a path coefficient analysis was performed. This method splits the correlation between a given trait and seed yield into two parts: (1) the direct effect of that trait on yield, and (2) indirect effects flowing through other related traits. In this analysis, seed yield per plant was treated as the outcome (dependent) variable, while all other characters were considered as potential contributors (independent variables). Additionally, a residual (unexplained) effect was estimated to assess how well the chosen set of traits accounted for variation in seed yield. A small residual value indicates that the traits included in the study explain most of the yield variability.

Residual Effect

A residual effect was estimated to determine how well the selected traits explained variation in seed yield. The residual value represents the contribution of other factors not included in the analysis. A small residual effect indicates that most of the variability in seed yield has been accounted for by the traits included in the study.

Results and Discussion

Correlation Patterns

The phenotypic and genotypic correlation coefficients among yield-related traits revealed meaningful interconnections. In almost all cases, genotypic correlations were larger than their phenotypic counterparts, indicating that environmental factors slightly masked the full expression of genetic associations. This pattern aligns with earlier observations by Patel et al. (2019) [16] . Seed yield per plant showed consistently strong and positive correlations with several component traits, evident at both the phenotypic and genotypic levels. Phenotypically, the highest positive association was with leaf area index (r = 0.790), followed closely by siliquae per plant (r = 0.786), biological yield (r = 0.677), test weight (r = 0.629), secondary branches (r = 0.611), seeds per siliqua (r = 0.396), harvest index (r = 0.341), main raceme length (r = 0.265), and primary branches (r = 0.217). Genotypically, the pattern was similar: leaf area index (r = 0.814), siliquae per plant (r = 0.797), biological yield (r = 0.675), secondary branches (r = 0.644), test weight (r = 0.642), seeds per siliqua (r = 0.411), harvest index (r = 0.350), main raceme length (r = 0.274), and primary branches (r = 0.245**). The fact that these relationships held across both levels suggests these traits are dependable indicators of seed yield and can be employed effectively in selection programs. Comparable findings have been reported by Prasad and Patil (2018) [19] , Nandi et al. (2021) [15] , and Chaudhary et al. (2023) [1] regarding positive associations of biological yield, primary branches, and seeds per siliqua with seed yield. In contrast, flowering-related traits displayed significant negative correlations with seed yield. Days to first flowering was inversely associated with yield (phenotypic r = -0.352, genotypic r = -0.360), and days to 50% flowering followed the same trend (phenotypic r = -0.311, genotypic r = -0.320). This implies that genotypes which flower earlier tend to achieve higher yields, possibly due to better alignment of reproductive growth with available resources. Similar negative links between flowering time and seed yield were noted by Meena et al. (2024) [14] . However, Lakra et al. (2020) [12] reported a positive association between days to 50% flowering and yield, highlighting that the phenologyyield relationship may shift depending on environment and genetic background. Days to maturity, meanwhile, showed only a very weak positive link with seed yield at both levels (phenotypic r = 0.005**, genotypic r = 0.007), indicating that maturity duration has almost no meaningful effect on productivity. This finding is partly supported by recent work in Indian mustard; for example, Perween et al. (2025) [18] found that days to maturity had a positive but relatively minor direct impact compared to major yield components like secondary branches and siliquae per plant. Similarly, Kumar et al. (2024) [7] observed a positive correlation but secondary importance for days to maturity.

Path Coefficient Results

Because correlation does not reveal whether a trait exerts its own direct influence or works through others, path analysis was performed at both the phenotypic and genotypic levels. As Dewey and Lu (1959) [3] noted, this method separates correlation coefficients into direct and indirect components. The residual effects were low, 0.0255 for phenotypic and 0.0145 for genotypic analysis, meaning the selected traits collectively explained most of the variation in seed yield. The smaller genotypic residual points to stronger genetic control over these traits.

Phenotypic Direct and Indirect Effects

Leaf area index (direct effect = 0.790) and number of siliquae per plant (0.786) had the largest positive direct effects on seed yield, highlighting the importance of photosynthetic capacity and reproductive sink size in determining yield. Other traits also showed notable direct contributions: biological yield (0.677), test weight (0.629), secondary branches (0.611), seeds per siliqua (0.396), harvest index (0.341), main raceme length (0.265), and primary branches (0.217). These results are in line with Meena et al. (2024), who concluded that biological yield and siliqua number are key direct drivers of seed yield in mustard. Days to first flowering (-0.352) and days to 50% flowering (-0.311) showed negative direct effects, reinforcing the advantage of early flowering. Oil content also had a slight negative direct effect (-0.120), suggesting that simultaneous improvement in yield and oil content may require careful, balanced selection. Similar negative relationships between delayed flowering and yield have been observed by Yadava et al. (2016) [22] , Patel et al. (2018) [19] . Additionally, leaf area index and siliquae per plant contributed indirectly through biological yield, further amplifying their total influence.

Genotypic Direct and Indirect Effects

The genotypic path analysis mirrored the phenotypic results but with slightly larger direct effect magnitudes, indicating strong genetic determination. Major direct contributors were leaf area index (0.814), siliquae per plant (0.797), biological yield (0.675), test weight (0.642), secondary branches (0.644), seeds per siliqua (0.411), harvest index (0.350), and main raceme length (0.274). The consistency across both levels increases confidence in using these traits for selection. As Kumar et al. (2024) [8] stated, siliquae per plant, biological yield, and test weight have strong direct effects and should be emphasized during selection, a conclusion fully supported here. Days to first flowering (-0.360) and days to 50% flowering (-0.320) again had negative direct effects at the genotypic level, confirming that early-maturing genotypes hold higher genetic potential for yield. Similar conclusions for days to 50% flowering were drawn by Lavanya et al. (2022) [13] , who stressed the value of early flowering in mustard breeding.

Conclusion

This study demonstrated that seed yield in Indian mustard is shaped by multiple interconnected morphological and physiological traits. Genotypic correlations were consistently stronger than phenotypic ones, confirming that most trait associations are under genetic control, though environment slightly dampens their visible expression. Correlation analysis identified leaf area index, siliquae per plant, biological yield, test weight, secondary branches, seeds per siliqua, harvest index, and main raceme length as having consistently strong positive links with seed yield at both levels. Conversely, days to first flower and days to 50% flower were negatively correlated with yield, favouring early-flowering types. Days to maturity had only a negligible association, suggesting little direct role in yield improvement. Path analysis further clarified these relationships by separating direct from indirect contributions. Leaf area index and siliquae per plant emerged as the most influential traits with the largest direct positive effects on seed yield. Biological yield, test weight, secondary branches, seeds per siliqua, and harvest index also made substantial direct contributions. The low residual effects at both levels confirmed that the included traits captured nearly all the variability in seed yield. Negative direct effects of flowering traits again emphasized the benefit of earliness. Taken together, the evidence from correlation and path analyses indicates that seed yield improvement can be achieved most effectively by selecting simultaneously for canopy development, branching capacity, reproductive sink size, and seed weight. These traits can serve as reliable selection criteria in breeding programs aimed at developing high-yielding Indian mustard varieties.

References

1. Chaudhary B, Lavanya GR, Yadav AP, Prajapati DK. Estimates of genetic variability, correlation and path analysis in Indian mustard (Brassica juncea L). J Oilseed Brassica. 2023;14(2):170-174.
2. Devi SA, Barua PK, Phukan A. Path analysis of yield traits in Brassica juncea. J Sci Res Rep. 2025;31(2):1-8.
3. Dewey DR, Lu KH. Correlation and path coefficient analysis of components of crested wheatgrass seed production. Agron J. 1959;57(3):515-518.
4. Food and Agriculture Organization of the United Nations. FAOSTAT: Food and Agriculture Organization Statistical Database. Rome: FAO; 2023.
5. Gadi J, Chakraborty NR, Imam Z. Correlation and path analysis for seed yield in Indian mustard. Plant Arch. 2025;25(1):203-210.
6. Gupta KK, Singh MK, Singh SM, Raghuvanshi AK. Genetic variability and path analysis in Indian mustard. J Exp Agric Int. 2025;47(6):27-33.
7. Kaur R, Kumar R, Kumar S, Aryan S. Character association studies in Indian mustard. Int J Biol Sci. 2025;7(8):146-153.
8. Kumar A, Singh YP, Dhangrah VK. Character association and gene action for yield and its contributing traits in Indian mustard (Brassica juncea L.). J Oilseeds Res. 2024;41(3-4):125-130.
9. Kumar A, Verma PN, Pandey VK, Saha S. Genetic variability, correlation, path coefficient analysis and diversity assessment of yield and yield related components in Indian mustard (Brassica juncea L.). Environ Ecol. 2024;42(4B):1897-1905.
10. Kumar S, Singh M, Yadav RK, Kumar S, Singh L, Shweta, et al. Assessment of correlation and path parameters in Indian mustard (Brassica juncea L. Czern. & Coss.). Int J Adv Biochem Res. 2025;9(11S):734742.
11. Kumar V, Sharma AK, Yogi RK, Singh VV. Rapeseedmustard cultivation in India: National overview and regional insights with reference to Madhya Pradesh. J Oilseed Brassica. 2025;16(2):149-157.
12. Lakra A, Tantuway G, Tirkey AE, Srivastava K. Genetic variability and trait association studies in Indian mustard (Brassica juncea L. Czern & Coss). Int J Curr Microbiol Appl Sci. 2020;9(1):2556-2563.
13. Lavanya K, Srikanth T, Padmaja D. Correlation and path coefficient analysis among yield and yield contributing traits in Indian mustard (Brassica juncea L.). Pharma Innov J. 2022;11(2):2926-2928.
14. Meena AK, Rajput SS, Yadav GL, Sheera A. Assessment of genetic variability, correlation and path analysis in Indian mustard (Brassica juncea L.). J Oilseed Brassica. 2024;15(2):246-251.
15. Nandi S, Rout S, Sur B, Chakraborty M, Hijam L, Roy SK, et al. Assessment of genetic parameters, trait International Journal of Agriculture and Food Science https://www.agriculturaljournals.com association and diversity analysis in Indian mustard (Brassica juncea). Biol Forum. 2021;13:95-109.
16. Patel P, Yadav K, Sharma N. Correlation and path analysis in Indian mustard for yield traits. J Oilseeds Res. 2018;35:250-255.
17. Patel JR, Prajapati KP, Patel PJ, Patel BK, Jat AL, Desai AG. Genetic variability and character association analysis for seed yield and its attributes in Indian mustard [Brassica juncea (L.) Czern and Coss]. Pharma Innov J. 2019;8(4):872-876.
18. Perween S, Kumar A, Chakraborty M, Mahto CS, Barnwal MK. Dissecting yield and yield-associated traits in Indian mustard (Brassica juncea L.): Insights from correlation and path analysis. J Oilseed Brassica. 2025;16(1):122-126.
19. Prasad G, Patil BR. Genetic variability and heritability studies for yield and attributes in Indian mustard. J Pharmacogn Phytochem. 2018;7:519-522.
20. Searle SR. Phenotypic, genotypic and environmental correlations. Biometrics. 1961;17:474-480.
21. United States Department of Agriculture, Foreign Agricultural Service. India: Oilseeds and Products Update. Washington (DC): USDA-FAS; 2024.
22. Yadava DK, Kumar S, Chauhan JS. Genetic studies on flowering traits and yield in mustard. Indian J Genet Plant Breed. 2016;76:94-99.